I’m going to be blunt: your approach to AI art is probably all wrong. You're wasting time and money on things that don't matter. As an investor in over 200 companies, including some of the foundational players in the AI space like OpenAI, Anthropic, and Scale AI, I see hundreds of pitch decks and product mockups. The same mistakes keep popping up.
It’s time for some tough love. That slick, hyper-detailed fantasy landscape you spent three hours prompting Midjourney to create? It might look cool, but it’s probably hurting your business more than it’s helping. You’re burning cash on GPU cycles and, more importantly, you’re burning your team’s time on a process that yields generic, forgettable results.
I saw a pitch deck last week from a promising fintech startup. The product was solid, the team was sharp, but the branding was a mess. Every image was a generic, glossy AI creation of abstract shapes and glowing lines. It looked like every other "futuristic" tech company out there. It told me nothing about their unique value. They were so focused on looking like the future that they forgot to build an identity. This isn’t just about aesthetics; it’s about your bottom line.
The Siren Song of the Perfect Prompt
One of the biggest traps I see founders fall into is the obsession with the "perfect prompt." They spend hours, even days, tweaking a string of text, adding arcane parameters, and chasing a photorealistic dragon that has exactly 17 scales on its left nostril. I get it. It feels like you’re mastering a new skill. It’s a fun puzzle.
But here’s the hard truth: prompt engineering is a dead-end job. The models are getting smarter. Soon, you’ll be able to just say, "a sad king in a crumbling castle," and the AI will understand the nuance, the mood, and the composition without a 300-word instruction manual. The value isn’t in the prompt; it’s in the idea.
Back when I was building RemoteTeam, which was later acquired by Gusto, we didn’t have these tools. We had to hire illustrators. The process was slow and expensive. We couldn’t just say "create a graphic of a happy remote worker." We had to have a conversation. We had to build a creative brief. We had to think about our brand, our message, our audience. The constraints forced us to be intentional. Today, the lack of constraints is making us lazy.
Your goal isn’t to become the world’s best prompt engineer. Your goal is to build a brand that connects with customers. Stop tweaking keywords and start thinking about strategy.
Your Generic Style is a Generic Brand
Another mistake is relying on the default styles of tools like Midjourney or DALL-E. You type in a prompt, and you get back an image that looks… exactly like every other image made with that tool. It’s the "Midjourney look." It’s clean, it’s detailed, but it’s also completely anonymous.
Using the default style is like building a restaurant and using the default Comic Sans font for your menu. It shows a lack of care. A lack of vision. Your visual identity is a core part of your brand. It’s the first thing a potential customer sees. Do you want their first impression to be "oh, another one of those"?
I advised a portfolio company last year that was struggling with user engagement on their app. Their interface was clean but sterile. The illustrations were all stock AI generations. We spent a weekend not writing prompts, but creating a visual identity guide. We looked at artists they admired, color palettes that evoked the right emotion, and compositional rules that would make their app feel unique. Then, and only then, did we start using AI. We didn’t use it to generate final images. We used it to rapidly prototype ideas within their new style guide. The final assets were a hybrid of AI generation and human refinement. Engagement shot up by 40% in the next quarter.
Art School for Robots: What You’re Missing
The real secret to great AI art isn’t about the AI. It’s about the art. The principles of composition, color theory, and visual storytelling haven’t changed. The tool is new, but the craft is ancient. You wouldn’t expect to write a great novel just because you have Microsoft Word, would you? The same applies here.
Here’s what the pros are doing instead of just mashing the "generate" button:
1. They Think Like a Director, Not a Prompter
Before you write a single word, you need a vision. What is the story of this image? What is the emotion? Who is the hero? What is the focal point? Instead of starting with "a futuristic cityscape," start with a purpose. "An image that shows our app brings order to the chaos of city life." See the difference? The first is a description. The second is a story.
Think about camera angles. Lighting. Lens choice. Is it a wide shot or a close-up? Is the lighting soft and optimistic or harsh and dramatic? These are the decisions that create compelling images, and most prompting amateurs completely ignore them.
2. They Build a "Style Bible"
Great brands have a consistent visual language. You need one too. Stop chasing random styles and build your own. A Style Bible is a document that defines your visual identity. It should include:
- Color Palette: A primary, secondary, and accent palette. Define the exact hex codes.
- Compositional Rules: Do you use a lot of negative space? Are your compositions symmetrical or asymmetrical? Do you follow the rule of thirds?
- Reference Artists: List 3-5 artists (photographers, painters, illustrators) whose work embodies your desired aesthetic. You can use their names in prompts to guide the AI.
- Texture and Detail: Do you prefer clean, flat graphics or rich, textured images? High detail or minimalist?
Once you have this, you can use it to create a "master prompt" or a set of style parameters that ensures everything you generate feels like it came from the same universe.
3. They Embrace the "Cyborg" Workflow
The best results I’ve seen don’t come straight out of the AI. They’re a collaboration between the machine and a human artist. This is the "Cyborg" approach. You use the AI for what it’s good at: rapid idea generation, creating base textures, and exploring variations. Then, you bring that raw material into a tool like Photoshop or Procreate for the final 20%.
A human artist can fix weird hands, adjust the composition, tweak the colors, and add those subtle details that sell the image as "real." This is where the magic happens. It’s not about replacing artists; it’s about giving them superpowers. One of my most successful investments, a gaming company, uses this method. Their art team can now produce 10x the number of assets they could before, and the quality is higher because they can iterate on more ideas.
Stop Wasting Money. Start Creating Value.
Look, I’m not saying you should stop using AI for art. That would be like telling a factory to stop using assembly lines. This technology is an incredible force multiplier. But you have to use it intelligently.
Stop thinking of it as a magic button and start thinking of it as a very powerful, very fast, and sometimes very dumb intern. It can do the legwork, but it needs your vision, your strategy, and your taste to create something truly valuable.
Your time is your most precious asset. Every hour you spend chasing the perfect AI prompt for a generic image is an hour you’re not spending on your product, your customers, or your strategy. Every dollar you spend on GPU time for images that don’t build your brand is a dollar wasted.
So, here’s my challenge to you. The next time you need a visual asset, don’t just open Midjourney and start typing. Take 30 minutes. Write a one-paragraph creative brief. Define the story, the emotion, and the purpose. Open your Style Bible. Then, and only then, start generating. And be prepared to take that output and make it your own.
That’s how you go from being a prompt monkey to a creative director. That’s how you stop wasting money and start building a brand that people remember. It requires more thought, but the return on that investment is immense. Don’t be lazy. Your startup’s future might depend on it.
Frequently Asked Questions
What's the most common pushback you get on this?
People often push back by citing exceptions or edge cases. And they're usually right that exceptions exist. But building a strategy around exceptions rather than patterns is a losing game for most founders.
How can I apply this thinking to my own situation?
Start by identifying the core principle behind the opinion, not the specific example. Then ask yourself: does this principle apply to my context? If yes, test it in a small, low-risk way before going all in.
How has this view evolved over time?
My thinking on most topics has changed significantly over the years. Early in my career, I held many conventional views that experience proved wrong. I try to update my beliefs when the evidence changes.